anonymousCML · Rocket-ZZW · Jan 4, 2022 · Jan 4, 2022 · Jan 4, 2022 · Jan 4, 2022
diff --git a/Params.py b/Params.py
diff --git a/README.md b/README.md
@@ -1,26 +1,90 @@
 # CML
-The implementation of Contrastive Meta Learning with Behavior Multiplicity for Recommendation
+ 
 
-mkdir /History
 
-mkdir /Model
+This repository contains PyTorch codes and datasets for the paper:
 
-### Requirements
-numpy==1.20.3
+> Wei, Wei and Huang, Chao and Xia, Lianghao and Xu, Yong and Zhao, Jiashu and Yin, Dawei. Contrastive Meta Learning with Behavior Multiplicity forRecommendation. <a href='https://arxiv.org/pdf/2202.08523.pdf'>Paper in arXiv</a>.
 
-scipy==1.6.2
 
-torch==1.8.1+cu111
+## Introduction
+Contrastive Meta Learning (CML) leverages multi-behavior learning paradigm to model diverse and multiplex user-item relationships, as well as tackling the label scarcity problem for target behaviors. The designed multi-behavior contrastive task is to capture the transferable user-item relationships from multi-typed user behavior data heterogeneity. And the proposed meta contrastive encoding scheme allows CML to preserve the personalized multi-behavior characteristics, so as to be reflective of the diverse behavior-aware user preference under a customized self-supervised framework.
 
-tqdm==4.61.2
 
+## Citation
+```
+@inproceedings{wei2022contrastive,
+  title={Contrastive meta learning with behavior multiplicity for recommendation},
+  author={Wei, Wei and Huang, Chao and Xia, Lianghao and Xu, Yong and Zhao, Jiashu and Yin, Dawei},
+  booktitle={Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining},
+  pages={1120--1128},
+  year={2022}
+}
+```
+
+
+## Environment
+
+The codes of CML are implemented and tested under the following development environment:
+
+- Python 3.6
+- torch==1.8.1+cu111
+- scipy==1.6.2
+- tqdm==4.61.2
+
+
+
+## Datasets
+
+#### Raw data：
+- IJCAI contest:  https://tianchi.aliyun.com/dataset/dataDetail?dataId=47
+- Retail Rocket: https://www.kaggle.com/retailrocket/ecommerce-dataset
+- Tmall:  https://tianchi.aliyun.com/dataset/dataDetail?dataId=649 
+#### Processed data：
+- The processed IJCAI are under the /datasets folder.
+
+
+## Usage
+
+The command to train CML on the Tmall/IJCAI/Retailrocket datasets are as follows. The commands specify the hyperparameter settings that generate the reported results in the paper.
+
+* Tmall
+```
+python main.py --path=./datasets/ --dataset=Tmall --opt_base_lr=1e-3 --opt_max_lr=5e-3 --opt_weight_decay=1e-4 --meta_opt_base_lr=1e-4 --meta_opt_max_lr=2e-3 --meta_opt_weight_decay=1e-4 --meta_lr=1e-3 --batch=8192 --meta_batch=128 --SSL_batch=18
+```
+* IJCAI
+```
+python main.py --path=./datasets/ --dataset=IJCAI_15 --sampNum=10 --opt_base_lr=1e-3 --opt_max_lr=2e-3 --opt_weight_decay=1e-4 --meta_opt_base_lr=1e-4 --meta_opt_max_lr=1e-3 --meta_opt_weight_decay=1e-4 --meta_lr=1e-3 --batch=8192 --meta_batch=128 --SSL_batch=30 
+```
+* Retailrocket
+```
+python main.py --path=./datasets/ --dataset='retailrocket' --sampNum=40 --lr=3e-4 --opt_base_lr=1e-4 --opt_max_lr=1e-3 --opt_weight_decay=1e-4 --opt_weight_decay=1e-4 --meta_opt_base_lr=1e-4 --meta_opt_max_lr=1e-3 --meta_opt_weight_decay=1e-3 --meta_lr=1e-3 --batch=2048 --meta_batch=128 --SSL_batch=15
+```         
+
+
+
+<!-- Important arguments:
+* `reg`: It is the weight for weight-decay regularization. We tune this hyperparameter from the set `{1e-2, 1e-3, 1e-4, 1e-5}`.
+* `ssl_reg`: This is the weight for the hypergraph-graph contrastive learning loss. The value is tuned from `1e-2` to `1e-8`.
+* `temp`: This is the temperature factor in the InfoNCE loss in our contrastive learning. The value is selected from `{10, 3, 1, 0.3, 0.1}`.
+* `keepRate`: It denotes the rate to keep edges in the graph dropout, which is tuned from `{0.25, 0.5, 0.75, 1.0}`.
+ -->
+
+
+
+
+
+<!-- ## Acknowledgement
+ -->
+
+
+
+
+
+It will be released again in few days in the optimized code version.
 
-### Datasets
-IJCAI contest:  https://tianchi.aliyun.com/dataset/dataDetail?dataId=47
 
-Retail Rocket: https://www.kaggle.com/retailrocket/ecommerce-dataset
 
-Tmall:  https://tianchi.aliyun.com/dataset/dataDetail?dataId=649 
 
 
 

diff --git a/Utils/README.md b/Utils/README.md
@@ -0,0 +1 @@
+.
diff --git a/__pycache__/BGNN.cpython-38.pyc b/__pycache__/BGNN.cpython-38.pyc
diff --git a/__pycache__/BSTRec.cpython-38.pyc b/__pycache__/BSTRec.cpython-38.pyc
diff --git a/__pycache__/BSTRec_beh.cpython-38.pyc b/__pycache__/BSTRec_beh.cpython-38.pyc
diff --git a/__pycache__/DataHandler.cpython-37.pyc b/__pycache__/DataHandler.cpython-37.pyc
diff --git a/__pycache__/DataHandler.cpython-38.pyc b/__pycache__/DataHandler.cpython-38.pyc
diff --git a/__pycache__/MB_GCN.cpython-38.pyc b/__pycache__/MB_GCN.cpython-38.pyc
diff --git a/__pycache__/MB_GCN_test.cpython-38.pyc b/__pycache__/MB_GCN_test.cpython-38.pyc
diff --git a/__pycache__/MB_GCN_test_behavior.cpython-38.pyc b/__pycache__/MB_GCN_test_behavior.cpython-38.pyc
diff --git a/__pycache__/MB_GCN_test_behavior_no_RNN.cpython-38.pyc b/__pycache__/MB_GCN_test_behavior_no_RNN.cpython-38.pyc
diff --git a/__pycache__/MB_GCN_test_behavior_normal_RNN.cpython-38.pyc b/__pycache__/MB_GCN_test_behavior_normal_RNN.cpython-38.pyc
diff --git a/__pycache__/MB_GCN_test_behavior_self_attention_knowledge.cpython-38.pyc b/__pycache__/MB_GCN_test_behavior_self_attention_knowledge.cpython-38.pyc
diff --git a/__pycache__/MV_Net.cpython-38.pyc b/__pycache__/MV_Net.cpython-38.pyc
diff --git a/__pycache__/MV_Net_beh.cpython-38.pyc b/__pycache__/MV_Net_beh.cpython-38.pyc
diff --git a/__pycache__/MV_Net_one_weights.cpython-38.pyc b/__pycache__/MV_Net_one_weights.cpython-38.pyc
diff --git a/__pycache__/NMTR.cpython-38.pyc b/__pycache__/NMTR.cpython-38.pyc
diff --git a/__pycache__/Params.cpython-37.pyc b/__pycache__/Params.cpython-37.pyc
diff --git a/__pycache__/Params.cpython-38.pyc b/__pycache__/Params.cpython-38.pyc
diff --git a/__pycache__/cell.cpython-38.pyc b/__pycache__/cell.cpython-38.pyc
diff --git a/__pycache__/cell_self_attention_knowledge.cpython-38.pyc b/__pycache__/cell_self_attention_knowledge.cpython-38.pyc
diff --git a/__pycache__/evaluate.cpython-37.pyc b/__pycache__/evaluate.cpython-37.pyc
diff --git a/__pycache__/evaluate.cpython-38.pyc b/__pycache__/evaluate.cpython-38.pyc
diff --git a/__pycache__/graph_utils.cpython-38.pyc b/__pycache__/graph_utils.cpython-38.pyc
diff --git a/__pycache__/graph_utils_gcn.cpython-38.pyc b/__pycache__/graph_utils_gcn.cpython-38.pyc
diff --git a/__pycache__/graph_utils_time.cpython-38.pyc b/__pycache__/graph_utils_time.cpython-38.pyc
diff --git a/__pycache__/hypergraph_utils.cpython-37.pyc b/__pycache__/hypergraph_utils.cpython-37.pyc
diff --git a/__pycache__/hypergraph_utils.cpython-38.pyc b/__pycache__/hypergraph_utils.cpython-38.pyc
diff --git a/__pycache__/meta_weight_net.cpython-38.pyc b/__pycache__/meta_weight_net.cpython-38.pyc
diff --git a/__pycache__/meta_weight_net_single.cpython-38.pyc b/__pycache__/meta_weight_net_single.cpython-38.pyc
diff --git a/__pycache__/meta_weight_net_trans.cpython-38.pyc b/__pycache__/meta_weight_net_trans.cpython-38.pyc
diff --git a/__pycache__/myModel.cpython-37.pyc b/__pycache__/myModel.cpython-37.pyc
diff --git a/__pycache__/myModel.cpython-38.pyc b/__pycache__/myModel.cpython-38.pyc
diff --git a/__pycache__/myModel_NoGNN.cpython-38.pyc b/__pycache__/myModel_NoGNN.cpython-38.pyc
diff --git a/__pycache__/myModel_NoGNN_mean.cpython-38.pyc b/__pycache__/myModel_NoGNN_mean.cpython-38.pyc
diff --git a/__pycache__/myModel_NoGNN_tmall.cpython-38.pyc b/__pycache__/myModel_NoGNN_tmall.cpython-38.pyc
diff --git a/__pycache__/myModel_RNN.cpython-38.pyc b/__pycache__/myModel_RNN.cpython-38.pyc
diff --git a/__pycache__/myModel_mean.cpython-38.pyc b/__pycache__/myModel_mean.cpython-38.pyc
diff --git a/__pycache__/myModel_mean_tmall.cpython-38.pyc b/__pycache__/myModel_mean_tmall.cpython-38.pyc
diff --git a/__pycache__/myModel_meta_RNN.cpython-38.pyc b/__pycache__/myModel_meta_RNN.cpython-38.pyc
diff --git a/__pycache__/myModel_meta_RNN_GAT.cpython-38.pyc b/__pycache__/myModel_meta_RNN_GAT.cpython-38.pyc
diff --git a/__pycache__/myModel_meta_RNN_GCN.cpython-38.pyc b/__pycache__/myModel_meta_RNN_GCN.cpython-38.pyc
diff --git a/__pycache__/myModel_meta_RNN_one_gru.cpython-38.pyc b/__pycache__/myModel_meta_RNN_one_gru.cpython-38.pyc
diff --git a/__pycache__/myModel_meta_RNN_one_gru_target_auxiliary.cpython-38.pyc b/__pycache__/myModel_meta_RNN_one_gru_target_auxiliary.cpython-38.pyc
diff --git a/__pycache__/myModel_multi_head.cpython-38.pyc b/__pycache__/myModel_multi_head.cpython-38.pyc
diff --git a/__pycache__/myModel_ori.cpython-38.pyc b/__pycache__/myModel_ori.cpython-38.pyc
diff --git a/data/README.md b/data/README.md
@@ -0,0 +1 @@
+.
diff --git a/main.py b/main.py
@@ -486,7 +486,8 @@ def trainEpoch(self):
         time = datetime.datetime.now()
         print("start_ng_samp:  ", time)
         train_loader.dataset.ng_sample()
-        time = datetime.datetime.now()
+
+
         print("end_ng_samp:  ", time)
 
         epoch_loss = 0